Also, if you are running at DB2 Release 8.1 or above, take a look at
whether your DB2 guy has specified page fixing for the DB2 buffer
pools. Improper selection of which DB2 buffer pools to page-fix can suck
up real memory and cause paging problems.
Don
******
Don Deese, Computer Management Sciences, Inc.
Voice: (804) 776-7109 Fax: (8043) 776-7139
http://www.cpexpert.org
******
At 11:16 AM 11/12/2008, you wrote:
How much memory do you have on the system? DB2 will page like crazy if
there is not sufficient. I have 14GB on my heavy DB2 Lpar and it never
consumes the paging datasets.
Prior to that I did not have sufficient real memory and got paged to death.
Lizette
>
>I have a situation with our page datasets that I fixed, at least for the
time being. I am curious though as to the behaviour of the page datasets
in relation to DB2.
>
>First, a little history of the problem. When I started here, one thing
I noticed was the page datasets were kind of small, and of different
sizes. We had 4 local page DS, the largest one being 1200 Cyl. This
past weekend, I made 4 full 3390 Mod 3 local page datasets, so I'm sure
the problems we had won't reoccur.
>
>I did notice, since I have been taking screen prints each day, that the
page dataset usage has been going up each day. On Monday, they averaged
8% full according to the RMF Mon 2 Page/Swap Data Set Activity
report. This morning, they are almost 20% full. I assume by Friday when
I go home they will be at least 30% full, and probably limiting block paging.
>
>My question is: is this normal DB2 behaviour? Or are we doing
something wrong. The guy who does some of the DB2 work thinks that DB2
caching of data is causing this, and that the problem is agravated
because most of the batch jobs that run at night use REGION=0M to get
maximum region size.
>
>Another thing I noticed while looking at the RMF Mon 3 Storage Frames
Report, the Production and the Quality regions have almost doubled the
number of Aux Slots held in the page datasets. P went from 72K to 128K,
and Q went from 66K to 117K. This was in a period of time of one day.
>
>The DB2 regions are started on Sunday, and left up all week until we IPL
Sunday morning. We still do weekly IPLs, just like my last full time job
at P&H!
>
>Anyway, is this behaviour normal? Are there certain things we might be
doing that we should be looking for?
>
>Thanks,
>
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html
